Setup

Necessary Libraries

library(MicrobeR)
library(dada2)
library(vegan)
library(ape)
library(philr)
library(lmerTest)
library(tidyverse)
library(readxl)
library(phyloseq)
library(ggtree)
library(qiime2R)
library(ALDEx2)
library(gghighlight)
library(ggpubr)
library(patchwork)
sessionInfo()
## R version 3.6.3 (2020-02-29)
## Platform: x86_64-apple-darwin15.6.0 (64-bit)
## Running under: macOS Mojave 10.14.6
## 
## Matrix products: default
## BLAS:   /Library/Frameworks/R.framework/Versions/3.6/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/3.6/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] stats     graphics  grDevices utils     datasets  methods   base     
## 
## other attached packages:
##  [1] patchwork_1.1.1   ggpubr_0.4.0      gghighlight_0.3.0 ALDEx2_1.18.0    
##  [5] qiime2R_0.99.34   ggtree_2.0.4      phyloseq_1.30.0   readxl_1.3.1     
##  [9] forcats_0.5.0     stringr_1.4.0     dplyr_0.8.5       purrr_0.3.4      
## [13] readr_1.3.1       tidyr_1.0.2       tibble_3.0.1      ggplot2_3.3.0    
## [17] tidyverse_1.3.0   lmerTest_3.1-2    lme4_1.1-23       Matrix_1.2-18    
## [21] philr_1.12.0      ape_5.3           vegan_2.5-6       lattice_0.20-38  
## [25] permute_0.9-5     dada2_1.14.1      Rcpp_1.0.4        MicrobeR_0.3.2   
## 
## loaded via a namespace (and not attached):
##   [1] backports_1.1.6             Hmisc_4.4-0                
##   [3] fastmatch_1.1-0             plyr_1.8.6                 
##   [5] igraph_1.2.5                lazyeval_0.2.2             
##   [7] splines_3.6.3               BiocParallel_1.20.1        
##   [9] GenomeInfoDb_1.22.1         rtk_0.2.5.8                
##  [11] digest_0.6.25               foreach_1.5.0              
##  [13] htmltools_0.5.0             fansi_0.4.1                
##  [15] checkmate_2.0.0             magrittr_1.5               
##  [17] memoise_1.1.0               cluster_2.1.0              
##  [19] DECIPHER_2.14.0             openxlsx_4.1.5             
##  [21] Biostrings_2.54.0           modelr_0.1.6               
##  [23] RcppParallel_5.0.0          matrixStats_0.56.0         
##  [25] jpeg_0.1-8.1                colorspace_1.4-1           
##  [27] blob_1.2.1                  rvest_0.3.5                
##  [29] haven_2.2.0                 xfun_0.13                  
##  [31] crayon_1.3.4                RCurl_1.98-1.2             
##  [33] jsonlite_1.6.1              survival_3.1-8             
##  [35] phangorn_2.5.5              iterators_1.0.12           
##  [37] glue_1.4.0                  gtable_0.3.0               
##  [39] zlibbioc_1.32.0             XVector_0.26.0             
##  [41] DelayedArray_0.12.3         car_3.0-8                  
##  [43] Rhdf5lib_1.8.0              BiocGenerics_0.32.0        
##  [45] abind_1.4-5                 scales_1.1.0               
##  [47] DBI_1.1.0                   rstatix_0.6.0              
##  [49] htmlTable_1.13.3            viridisLite_0.3.0          
##  [51] tidytree_0.3.3              foreign_0.8-75             
##  [53] bit_1.1-15.2                Formula_1.2-3              
##  [55] stats4_3.6.3                DT_0.13                    
##  [57] truncnorm_1.0-8             htmlwidgets_1.5.1          
##  [59] httr_1.4.1                  RColorBrewer_1.1-2         
##  [61] acepack_1.4.1               ellipsis_0.3.0             
##  [63] pkgconfig_2.0.3             NADA_1.6-1.1               
##  [65] nnet_7.3-12                 dbplyr_1.4.3               
##  [67] tidyselect_1.0.0            rlang_0.4.5                
##  [69] reshape2_1.4.4              munsell_0.5.0              
##  [71] cellranger_1.1.0            tools_3.6.3                
##  [73] cli_2.0.2                   generics_0.0.2             
##  [75] RSQLite_2.2.0               ade4_1.7-15                
##  [77] broom_0.5.6                 evaluate_0.14              
##  [79] biomformat_1.14.0           yaml_2.2.1                 
##  [81] knitr_1.28                  bit64_0.9-7                
##  [83] fs_1.4.1                    zip_2.0.4                  
##  [85] nlme_3.1-144                xml2_1.3.2                 
##  [87] rstudioapi_0.11             compiler_3.6.3             
##  [89] curl_4.3                    plotly_4.9.2.1             
##  [91] png_0.1-7                   ggsignif_0.6.0             
##  [93] zCompositions_1.3.4         reprex_0.3.0               
##  [95] treeio_1.10.0               statmod_1.4.34             
##  [97] stringi_1.4.6               nloptr_1.2.2.1             
##  [99] multtest_2.42.0             vctrs_0.2.4                
## [101] pillar_1.4.3                lifecycle_0.2.0            
## [103] BiocManager_1.30.10         data.table_1.12.8          
## [105] bitops_1.0-6                GenomicRanges_1.38.0       
## [107] R6_2.4.1                    latticeExtra_0.6-29        
## [109] hwriter_1.3.2               ShortRead_1.44.3           
## [111] rio_0.5.16                  gridExtra_2.3              
## [113] IRanges_2.20.2              codetools_0.2-16           
## [115] boot_1.3-24                 MASS_7.3-51.5              
## [117] assertthat_0.2.1            picante_1.8.1              
## [119] rhdf5_2.30.1                SummarizedExperiment_1.16.1
## [121] withr_2.2.0                 GenomicAlignments_1.22.1   
## [123] Rsamtools_2.2.3             S4Vectors_0.24.4           
## [125] GenomeInfoDbData_1.2.2      mgcv_1.8-31                
## [127] parallel_3.6.3              hms_0.5.3                  
## [129] rpart_4.1-15                quadprog_1.5-8             
## [131] grid_3.6.3                  minqa_1.2.4                
## [133] rmarkdown_2.1               rvcheck_0.1.8              
## [135] carData_3.0-4               base64enc_0.1-3            
## [137] numDeriv_2016.8-1.1         Biobase_2.46.0             
## [139] lubridate_1.7.8

Theme

Whitecolor="#E69F00"
Chinesecolor="#0072B2"

# theme for pcoas
theme_pcoa<- function () { 
  theme_classic(base_size=10, base_family="Helvetica") +
  theme(axis.text = element_text(size=8, color = "black"), 
        axis.title = element_text(size=10, color="black"), 
        legend.text = element_text(size=8, color = "black"), 
        legend.title = element_text(size=10, color = "black"), 
        plot.title = element_text(size=10, color="black")) +
  theme(panel.border = element_rect(color="black", size=1, fill=NA))
}

# theme for boxplots
theme_boxplot<- function () { 
  theme_classic(base_size=10, base_family="Helvetica") +
  theme(axis.text.x = element_text(size=10, color = "black"),
        axis.text.y = element_text(size=8, color="black"),
        axis.title.x= element_blank(),
        axis.title.y = element_text(size=10, color="black"), 
        legend.position = "none")
}

Data Import

metadata<-read_excel("/Volumes/turnbaughlab/qb3share/qiyanang/IDEO_manuscript_Rd2/data/metadata_IDEO_cohort.xlsx") %>% column_to_rownames("SampleID")

SVtab<-read_qza("/Volumes/turnbaughlab/qb3share/qiyanang/16S_IDEO_FonlyNewPipeline/Output/ASV_table.qza")$data %>% as.data.frame() 

SVseq<-read_qza("/Volumes/turnbaughlab/qb3share/qiyanang/16S_IDEO_FonlyNewPipeline/Output/ASV_sequences.qza")$data %>% as.data.frame() %>% rename("SV"=x)

taxonomy<-read.delim("/Volumes/turnbaughlab/qb3share/qiyanang/16S_IDEO_FonlyNewPipeline/Output/ASV_d2taxonomy.txt", header=T)

lookup<-(SVseq %>% rownames_to_column("ASV")) %>%
  left_join(taxonomy, by="ASV") %>%
  column_to_rownames("ASV")
## Warning: Column `ASV` joining character vector and factor, coercing into
## character vector
# Tree
tree<-read_qza("/Volumes/turnbaughlab/qb3share/qiyanang/16S_IDEO_FonlyNewPipeline/Output/ASV_denovotree.qza")$data

Filter SV table

# Correct typos in SVtab colnames
names(SVtab) <- gsub(x = names(SVtab), pattern = "0B0", replacement = "OB0")  

# Subset SVtab to contain only White and Chinese samples
SVtab<-SVtab[,rownames(metadata)]
histogram(colSums(SVtab))

print("Read count for samples used in downstream analysis range 72091-319590")
## [1] "Read count for samples used in downstream analysis range 72091-319590"

Quality Filter

SVtab<-Confidence.Filter(SVtab, MINSAMPS = 2, MINREADS=10, VERBOSE=TRUE)
lookup<-lookup[rownames(SVtab),]
tree<-drop.tip(tree, tree$tip.label[!tree$tip.label %in% rownames(lookup)])

Normalized Tables

PHILR<-philr(
            t(SVtab+1), 
            tree, 
            part.weights='enorm.x.gm.counts', 
            ilr.weights='blw.sqrt'
            )
## Building Sequential Binary Partition from Tree...
## Building Contrast Matrix...
## Transforming the Data...
## Calculating ILR Weights...
## Warning in calculate.blw(tree, method = "sum.children"): Note: a total of 105
## tip edges with zero length have been replaced with a small pseudocount of the
## minimum non-zero edge length ( 5e-09 ).

Alpha Diversity

alphadiv <- data.frame(
  Shannon = vegan::diversity(Subsample.Table(SVtab), index = "shannon", MARGIN = 2), 
  FaithsPD = picante::pd(t(Subsample.Table(SVtab)), tree, include.root = F)$PD,
  Richness = specnumber(Subsample.Table(SVtab), MARGIN = 2)) %>% #Calc richness on subsampled table
  rownames_to_column("SampleID") %>%
  left_join(metadata %>% rownames_to_column("SampleID")) %>%
  select (SampleID, Ethnicity, IDEO_BMI_Class, Shannon, FaithsPD, Richness, BMI,`%BF`) %>%
  pivot_longer(cols=Shannon:Richness, names_to="alpha_metric")
## Subsampling feature table to 69340 , currently has  1308  taxa.
## ...sampled to 69340 reads with 1305 taxa
## Subsampling feature table to 69340 , currently has  1308  taxa.
## ...sampled to 69340 reads with 1305 taxa
## Subsampling feature table to 69340 , currently has  1308  taxa.
## ...sampled to 69340 reads with 1305 taxa
## Joining, by = "SampleID"
# plot alpha div metrics
alphadiv %>%
  ggplot(aes(x=IDEO_BMI_Class, y=value, fill=Ethnicity)) + 
  geom_boxplot(outlier.shape=NA) +
  facet_wrap(~alpha_metric, scales="free", nrow=1) +
  theme_boxplot() +
  theme(legend.position = "right") +
  scale_fill_manual(values = c(EA=Chinesecolor, W=Whitecolor)) +
  ylab("Alpha diversity") +
  stat_compare_means(method = "wilcox.test", paired = FALSE, label = "p.format")

#ggsave("/Volumes/turnbaughlab/qb3share/qiyanang/IDEO_manuscript_Rd2/figures/figures_16S_human_leanobese/alphadiv_leanobese.pdf", height=2.5, width=6, useDingbats=F)

# stats
alphadiv %>%
  group_by(alpha_metric, IDEO_BMI_Class) %>%
  do(
    broom::glance(wilcox.test(value~Ethnicity, data=., paired=F))
  ) %>%
  ungroup() -> results.alpha
## Warning in wilcox.test.default(x = c(268, 300, 281, 160, 202, 262, 164, : cannot
## compute exact p-value with ties
## Warning in wilcox.test.default(x = c(271, 147, 196, 181, 230, 169, 185, : cannot
## compute exact p-value with ties
Nice.Table(results.alpha)

Correlation bw alpha div and metabolic parameters

Correlation stats

alphadiv %>%
  group_by(alpha_metric, Ethnicity) %>%
  do(
    broom::glance(cor.test(~value+BMI, data=., method="spearman", conf.level = 0.95))
  ) %>%
  ungroup() -> results.alpha.cor.BMI
## Warning in cor.test.default(x = c(17.433724333, 10.718564998, 13.875079969, :
## Cannot compute exact p-value with ties
## Warning in cor.test.default(x = c(15.473810632, 12.618487713, 16.065054884, :
## Cannot compute exact p-value with ties
## Warning in cor.test.default(x = c(271, 147, 196, 181, 268, 300, 281, 160, :
## Cannot compute exact p-value with ties
## Warning in cor.test.default(x = c(237, 205, 237, 227, 216, 291, 151, 225, :
## Cannot compute exact p-value with ties
## Warning in cor.test.default(x = c(3.75391221497536, 3.28392747785931,
## 3.65868347495839, : Cannot compute exact p-value with ties
## Warning in cor.test.default(x = c(3.47753869036769, 4.01093947025265,
## 3.93037963333439, : Cannot compute exact p-value with ties
Nice.Table(results.alpha.cor.BMI)
alphadiv %>%
  group_by(alpha_metric, Ethnicity) %>%
  do(
    broom::glance(cor.test(~value+as.numeric(`%BF`), data=., method="spearman", conf.level = 0.95))
  ) %>%
  ungroup() -> results.alpha.cor.BF
## Warning in cor.test.default(x = c(17.433724333, 10.718564998, 13.875079969, :
## Cannot compute exact p-value with ties
## Warning in eval(predvars, data, env): NAs introduced by coercion
## Warning in cor.test.default(x = c(15.473810632, 12.618487713, 15.3622618, :
## Cannot compute exact p-value with ties
## Warning in cor.test.default(x = c(271, 147, 196, 181, 268, 300, 281, 160, :
## Cannot compute exact p-value with ties
## Warning in eval(predvars, data, env): NAs introduced by coercion
## Warning in cor.test.default(x = c(237, 205, 227, 216, 291, 151, 225, 203, :
## Cannot compute exact p-value with ties
## Warning in cor.test.default(x = c(3.75391221497536, 3.28392747785931,
## 3.65868347495839, : Cannot compute exact p-value with ties
## Warning in eval(predvars, data, env): NAs introduced by coercion
## Warning in cor.test.default(x = c(3.47753869036769, 4.01093947025265,
## 3.77358789674242, : Cannot compute exact p-value with ties
Nice.Table(results.alpha.cor.BF)

Plot Richness corr to BMI/BF (main fig)

# plot Richness to BMI
richness_bmi <- alphadiv %>%
  filter(alpha_metric=="Richness") %>%
  ggplot(aes(x=as.numeric(BMI), y=value, fill=Ethnicity)) +
  stat_smooth(method="lm", color="black", size=1) +
  geom_point(size=2, shape=21) +
  facet_wrap(~Ethnicity) +
  scale_fill_manual(values = c(EA=Chinesecolor, W=Whitecolor)) +
  theme_pcoa() +
  theme(legend.position = "none") +
  xlab("BMI") +
  ylab("Richness")

# plot Richness to %BF
richness_vat <- alphadiv %>%
  filter(alpha_metric=="Richness") %>%
  ggplot(aes(x=as.numeric(`%BF`), y=value, fill=Ethnicity)) +
  stat_smooth(method="lm", color="black", size=1) +
  geom_point(size=2, shape=21) +
  facet_wrap(~Ethnicity) +
  theme(legend.position = "none") +
  scale_fill_manual(values = c(EA=Chinesecolor, W=Whitecolor)) +
  theme_pcoa() +
  theme(legend.position = "none") +
  xlab("Body fat, %") +
  ylab("Richness")

# combine both plots
richness_bmi / richness_vat
## `geom_smooth()` using formula 'y ~ x'
## Warning in FUN(X[[i]], ...): NAs introduced by coercion

## Warning in FUN(X[[i]], ...): NAs introduced by coercion

## Warning in FUN(X[[i]], ...): NAs introduced by coercion
## `geom_smooth()` using formula 'y ~ x'
## Warning: Removed 1 rows containing non-finite values (stat_smooth).
## Warning: Removed 1 rows containing missing values (geom_point).

#ggsave("/Volumes/turnbaughlab/qb3share/qiyanang/IDEO_manuscript_Rd2/figures/figures_16S_human_leanobese/alphadiv_metabolic_corr_main.pdf", height=4, width=3.5, useDingbats=F)

Plot FaithsPD and Shannon (suppl fig)

# plot Shannon to BMI
shannon_bmi <- alphadiv %>%
  filter(alpha_metric=="Shannon") %>%
  ggplot(aes(x=as.numeric(BMI), y=value, fill=Ethnicity)) +
  stat_smooth(method="lm", color="black", size=1) +
  geom_point(size=2, shape=21) +
  facet_wrap(~Ethnicity) +
  scale_fill_manual(values = c(EA=Chinesecolor, W=Whitecolor)) +
  theme_pcoa() +
  theme(legend.position = "none") +
  xlab("BMI") +
  ylab("Shannon diversity")

# plot Shannon to %BF
shannon_vat <- alphadiv %>%
  filter(alpha_metric=="Shannon") %>%
  ggplot(aes(x=as.numeric(`%BF`), y=value, fill=Ethnicity)) +
  stat_smooth(method="lm", color="black", size=1) +
  geom_point(size=2, shape=21) +
  facet_wrap(~Ethnicity) +
  scale_fill_manual(values = c(EA=Chinesecolor, W=Whitecolor)) +
  theme_pcoa() +    
  theme(legend.position = "none") +
  xlab("Body fat, %") +
  ylab("Shannon diversity")

# plot FaithsPD to BMI
faiths_bmi <- alphadiv %>%
  filter(alpha_metric=="FaithsPD") %>%
  ggplot(aes(x=as.numeric(BMI), y=value, fill=Ethnicity)) +
  stat_smooth(method="lm", color="black", size=1) +
  geom_point(size=2, shape=21) +
  facet_wrap(~Ethnicity) +
  scale_fill_manual(values = c(EA=Chinesecolor, W=Whitecolor)) +
  theme_pcoa() +
  theme(legend.position = "none") +
  xlab("BMI") +
  ylab("Faith's diversity")

# plot FatihsPD to %BF
faiths_vat <- alphadiv %>%
  filter(alpha_metric=="FaithsPD") %>%
  ggplot(aes(x=as.numeric(`%BF`), y=value, fill=Ethnicity)) +
  stat_smooth(method="lm", color="black", size=1) +
  geom_point(size=2, shape=21) +
  facet_wrap(~Ethnicity) +
  scale_fill_manual(values = c(EA=Chinesecolor, W=Whitecolor)) +
  theme_pcoa() +
  theme(legend.position = "none") +
  xlab("Body fat, %") +
  ylab("Faith's diversity")

# Combine panels
(faiths_bmi | faiths_vat) / (shannon_bmi | shannon_vat) + plot_annotation(tag_levels = "A")
## `geom_smooth()` using formula 'y ~ x'
## Warning in FUN(X[[i]], ...): NAs introduced by coercion

## Warning in FUN(X[[i]], ...): NAs introduced by coercion

## Warning in FUN(X[[i]], ...): NAs introduced by coercion
## `geom_smooth()` using formula 'y ~ x'
## Warning: Removed 1 rows containing non-finite values (stat_smooth).
## Warning: Removed 1 rows containing missing values (geom_point).
## `geom_smooth()` using formula 'y ~ x'
## Warning in FUN(X[[i]], ...): NAs introduced by coercion
## Warning in FUN(X[[i]], ...): NAs introduced by coercion

## Warning in FUN(X[[i]], ...): NAs introduced by coercion
## `geom_smooth()` using formula 'y ~ x'
## Warning: Removed 1 rows containing non-finite values (stat_smooth).
## Warning: Removed 1 rows containing missing values (geom_point).

#ggsave("/Volumes/turnbaughlab/qb3share/qiyanang/IDEO_manuscript_Rd2/figures/figures_16S_human_leanobese/alphadiv_metabolic_corr_suppl.pdf", height=4, width=6, useDingbats=F)

Phylum Abundances

Summarize.Taxa(SVtab, lookup)$Phylum %>%  
  Make.CLR() %>%
  as.data.frame() %>%
  rownames_to_column("Feature") %>%
  gather(-Feature, key="SampleID", value="Abundance") %>% 
  left_join(metadata %>% rownames_to_column("SampleID"), by="SampleID") %>%
  select(Feature, SampleID, Abundance, SubjectID, Ethnicity, IDEO_BMI_Class) -> data
## WARNING: CLR being applied with relatively few features.
# top 5 phyla
# data %>%
#   group_by(Feature) %>%
#   summarize(AvgAbundance=mean(Abundance)) %>%
#   top_n(n=6, wt=AvgAbundance) -> top_phyla

# significant phyla
data %>%
  group_by(Feature) %>%
  do(
    broom::glance(wilcox.test(Abundance~Ethnicity, data=., paired=FALSE))
  ) %>%
  ungroup() -> results.phylum
results.phylum %>% filter(p.value<0.05) -> sig_phyla
Nice.Table(results.phylum)
# plot phyla
data %>%
  filter(Feature %in% sig_phyla$Feature) %>%
  separate(Feature, sep=";", into=c("Kingdom","Phylum")) %>%
  mutate(Phylum=factor(Phylum, levels=c("Firmicutes","Bacteroidota","Actinobacteriota","Proteobacteria","Verrucomicrobiota","Fusobacteriota"))) %>%
  ggplot(aes(x=IDEO_BMI_Class, y=Abundance, fill=Ethnicity)) +
  geom_boxplot(outlier.shape=NA) +  
  facet_wrap(~Phylum, nrow=1, scales="free") +
  theme_boxplot() +
  theme(legend.position = "right") +
  scale_fill_manual(values = c(EA=Chinesecolor, W=Whitecolor)) +
  ylab("Abundance (CLR)") +
  stat_compare_means(method = "wilcox.test",paired = FALSE,label = "p.format")

#ggsave("/Volumes/turnbaughlab/qb3share/qiyanang/IDEO_manuscript_Rd2/figures/figures_16S_human_leanobese/phyla_leanobese.pdf", height=2.5, width=10, useDingbats=F)

Beta Diversity (PhILR): Lean

# filter lean subjects
f.meta<-metadata %>% rownames_to_column("SampleID") %>% filter(IDEO_BMI_Class=="Lean")
f.PHILR<-PHILR[f.meta$SampleID,]

#Adonis
adonis<-vegan::adonis(dist(f.PHILR, method="euclidean") ~ Ethnicity, data=f.meta, permutations=10000)
adonis$aov.tab
## Permutation: free
## Number of permutations: 10000
## 
## Terms added sequentially (first to last)
## 
##           Df SumsOfSqs MeanSqs F.Model      R2 Pr(>F)    
## Ethnicity  1    3242.6  3242.6  2.9102 0.12171  3e-04 ***
## Residuals 21   23398.5  1114.2         0.87829           
## Total     22   26641.1                 1.00000           
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
#Calculate PCo axis values
euclid<-ape::pcoa(dist(f.PHILR, method="euclidean"))
var.PCo1 <- format(100*(euclid$values$Eigenvalues/sum(euclid$values$Eigenvalues))[1], digits=2, nsmall=1)
var.PCo2 <- format(100*(euclid$values$Eigenvalues/sum(euclid$values$Eigenvalues))[2], digits=2, nsmall=1)

#Plot pcoa
euclid$vectors %>%
  as.data.frame() %>%
  rownames_to_column("SampleID") %>%
  left_join(metadata %>% rownames_to_column("SampleID")) %>%
  ggplot(aes(x=Axis.1, y=Axis.2, fill=Ethnicity)) +
  geom_point(size=2, shape=21) +
  theme_pcoa() +
  ylab(paste0("PCo2 [",var.PCo2,"%]")) +
  xlab(paste0("PCo1 [",var.PCo1,"%]")) +
  scale_fill_manual(values = c(EA=Chinesecolor, W=Whitecolor)) +
  ggtitle(paste0("PhILR (Lean): p=",adonis$aov.tab$`Pr(>F)`,", r2=",round(adonis$aov.tab$R2[1],digits=3))) +
  theme(legend.position = "none")
## Joining, by = "SampleID"

#ggsave("/Volumes/turnbaughlab/qb3share/qiyanang/IDEO_manuscript_Rd2/figures/figures_16S_human_leanobese/pcoa_lean.pdf", height=2.5, width=2.5, useDingbats=F)

Beta Diversity (PhILR Euclidean): Obese

# filter obese subjects
f.meta<-metadata %>% rownames_to_column("SampleID") %>% filter(IDEO_BMI_Class=="Obese")
f.PHILR<-PHILR[f.meta$SampleID,]

#Adonis
adonis<-vegan::adonis(dist(f.PHILR, method="euclidean") ~ Ethnicity, data=f.meta, permutations=10000)
adonis$aov.tab
## Permutation: free
## Number of permutations: 10000
## 
## Terms added sequentially (first to last)
## 
##           Df SumsOfSqs MeanSqs F.Model      R2 Pr(>F)
## Ethnicity  1    1784.3  1784.3  1.3279 0.05947 0.1423
## Residuals 21   28217.0  1343.7         0.94053       
## Total     22   30001.3                 1.00000
#Calculate PCo axis values
euclid<-ape::pcoa(dist(f.PHILR, method="euclidean"))
var.PCo1 <- format(100*(euclid$values$Eigenvalues/sum(euclid$values$Eigenvalues))[1], digits=2, nsmall=1)
var.PCo2 <- format(100*(euclid$values$Eigenvalues/sum(euclid$values$Eigenvalues))[2], digits=2, nsmall=1)

#Plot pcoa
euclid$vectors %>%
  as.data.frame() %>%
  rownames_to_column("SampleID") %>%
  left_join(metadata %>% rownames_to_column("SampleID")) %>%
  ggplot(aes(x=Axis.1, y=Axis.2, fill=Ethnicity)) +
  geom_point(size=2, shape=21) +
  theme_pcoa() +
  ylab(paste0("PCo2 [",var.PCo2,"%]")) +
  xlab(paste0("PCo1 [",var.PCo1,"%]")) +
  scale_fill_manual(values = c(EA=Chinesecolor, W=Whitecolor)) +
  ggtitle(paste0("PhILR (Obese): p=",adonis$aov.tab$`Pr(>F)`,", r2=",round(adonis$aov.tab$R2[1],digits=3))) +
  theme(legend.position = "none")
## Joining, by = "SampleID"

#ggsave("/Volumes/turnbaughlab/qb3share/qiyanang/IDEO_manuscript_Rd2/figures/figures_16S_human_leanobese/pcoa_obese.pdf", height=2.5, width=2.5, useDingbats=F)

Differential Abundance on genera: Aldex2 (Lean subjects)

# filter lean subjects
f.meta<-metadata %>% rownames_to_column("SampleID") %>% filter(IDEO_BMI_Class=="Lean")
f.SVtab<-SVtab[,f.meta$SampleID]

# summarize table to genus level
genera<-Summarize.Taxa(f.SVtab, lookup)$Genus
f.genera<-Fraction.Filter(genera,0.0005)
## [1] "Filtering table at a min fraction of 5e-04 of feature table..."
## [1] "...There are 5289790 reads and 243  features"
## [1] "...After filtering there are 5220143 reads and 99 OTUs"
# aldex
results <- aldex(f.genera, f.meta$Ethnicity, mc.samples = 128, denom = "all", test = "t", effect = TRUE, include.sample.summary = TRUE)
## aldex.clr: generating Monte-Carlo instances and clr values
## operating in serial mode
## computing center with all features
## aldex.ttest: doing t-test
## aldex.effect: calculating effect sizes
results_lean <-
results %>%
  rownames_to_column("Feature") %>%
  select(Feature, 
         logFC_Between=diff.btw, 
         logFC_Within=diff.win,
         Abundance_EA=rab.win.EA,
         Abundance_W=rab.win.W,
         Pvalue=we.ep, 
         FDR=we.eBH, 
         EffectSize=effect) %>%
  mutate(logFC_EA_vs_W=-(logFC_Between)) %>%
  separate(Feature,sep=";",into=c("K","P","C","O","F","Genus"),remove=F)

# significant results
sigres_lean <-
results_lean %>%
  filter(FDR<0.1 & abs(logFC_Between)>1)

Nice.Table(sigres_lean)
# volcano plot
ggplot(results_lean, aes(x = logFC_EA_vs_W, y=-log10(FDR))) + 
  geom_point() + 
  gghighlight(FDR < 0.1 & abs(logFC_Between)>1, label_key = Genus) +
  theme_bw() +
  theme(plot.title = element_text(hjust = 0.5)) +
  ggtitle("Lean individuals") +
  xlab("Log2 fold difference (EA/W)") +
  ylab("-Log10(FDR)") +
  xlim(-6,6)
## Warning: Removed 1 rows containing missing values (geom_point).

#ggsave("/Volumes/turnbaughlab/qb3share/qiyanang/IDEO_manuscript_Rd2/figures/figures_16S_human_leanobese/aldex_lean_volplot.pdf", height=2.5, width=2.5, useDingbats=F)

# plot abundances of significant genera
genera %>%
  Make.CLR() %>%
  as.data.frame() %>%
  rownames_to_column("Feature") %>%
  filter(Feature %in% sigres_lean$Feature) %>%
  pivot_longer(-Feature, names_to = "SampleID", values_to = "CLR") %>%
  left_join(f.meta) %>%
  separate(Feature, sep=";", into=c("K","P","C","O","Family","Genus")) %>%
  ggplot(aes(x=Ethnicity, y=CLR, fill=Ethnicity)) +
  geom_boxplot(outlier.shape=NA) +
  geom_jitter(shape=21, size=1, height=0, width=0.1) +
  facet_wrap(~Genus, scales="free", nrow=2) +
  theme_boxplot() +
  scale_fill_manual(values = c(EA=Chinesecolor, W=Whitecolor)) +
  ylab("Abundance (CLR)")
## Joining, by = "SampleID"

#ggsave("/Volumes/turnbaughlab/qb3share/qiyanang/IDEO_manuscript_Rd2/figures/figures_16S_human_leanobese/aldex_lean_genera.pdf", height=4, width=6, useDingbats=F)

Differential Abundance on genera: Aldex2 (Obese subjects)

# filter obese subjects
f.meta<-metadata %>% rownames_to_column("SampleID") %>% filter(IDEO_BMI_Class=="Obese")
f.SVtab<-SVtab[,f.meta$SampleID]

# summarize table to genus level
genera<-Summarize.Taxa(f.SVtab, lookup)$Genus
f.genera<-Fraction.Filter(genera,0.0005)
## [1] "Filtering table at a min fraction of 5e-04 of feature table..."
## [1] "...There are 4628776 reads and 243  features"
## [1] "...After filtering there are 4560655 reads and 94 OTUs"
# aldex
results <- aldex(f.genera, f.meta$Ethnicity, mc.samples = 128, denom = "all", test = "t", effect = TRUE, include.sample.summary = TRUE)
## aldex.clr: generating Monte-Carlo instances and clr values
## operating in serial mode
## computing center with all features
## aldex.ttest: doing t-test
## aldex.effect: calculating effect sizes
results_obese <-
results %>%
  rownames_to_column("Feature") %>%
  select(Feature, 
         logFC_Between=diff.btw, 
         logFC_Within=diff.win,
         Abundance_EA=rab.win.EA,
         Abundance_W=rab.win.W,
         Pvalue=we.ep, 
         FDR=we.eBH, 
         EffectSize=effect) %>%
  mutate(logFC_EA_vs_W=-(logFC_Between)) %>%
  separate(Feature,sep=";",into=c("K","P","C","O","F","Genus"),remove=F)

# significant results
sigres_obese <-
results_obese %>%
  filter(FDR<0.1 & abs(logFC_Between)>1)
Nice.Table(sigres_obese)
# volcano plot
ggplot(results_obese, aes(x = logFC_EA_vs_W, y=-log10(FDR))) + 
  geom_point() + 
  gghighlight(FDR < 0.1 & abs(logFC_Between)>1, label_key = Genus) +
  theme_bw() +
  theme(plot.title = element_text(hjust = 0.5)) +
  ggtitle("Obese individuals") +
  xlab("Log2 fold difference (EA/W)") +
  ylab("-Log10(FDR)") +
  xlim(-9,9)

#ggsave("/Volumes/turnbaughlab/qb3share/qiyanang/IDEO_manuscript_Rd2/figures/figures_16S_human_leanobese/aldex_obese_volplot.pdf", height=2.5, width=2.5, useDingbats=F)

# plot abundances of significant genera
genera %>%
  Make.CLR() %>%
  as.data.frame() %>%
  rownames_to_column("Feature") %>%
  filter(Feature %in% sigres_obese$Feature) %>%
  pivot_longer(-Feature, names_to = "SampleID", values_to = "CLR") %>%
  left_join(f.meta) %>%
  separate(Feature, sep=";", into=c("K","P","C","O","Family","Genus")) %>%
  ggplot(aes(x=Ethnicity, y=CLR, fill=Ethnicity)) +
  geom_boxplot(outlier.shape=NA) +
  geom_jitter(shape=21, size=1, height=0, width=0.1) +
  facet_wrap(~Genus, scales="free", nrow=1) +
  theme_boxplot() +
  scale_fill_manual(values = c(EA=Chinesecolor, W=Whitecolor)) +
  ylab("Abundance (CLR)")
## Joining, by = "SampleID"

#ggsave("/Volumes/turnbaughlab/qb3share/qiyanang/IDEO_manuscript_Rd2/figures/figures_16S_human_leanobese/aldex_obese_genera.pdf", height=2, width=4, useDingbats=F)